Computer-executed content recommendation engines have been provided that provide a user with recommended content based on user behavior. As understood herein, in the TV context a recommendation engine may correlate viewed channels, program times, and program types, both past and current, to other content that matches the user behavior within the constraints of the engine. The recommendation engine then outputs a list of recommended content. Present principles understand that in the TV context, the recommended content list advantageously should be integrated with the overall look and feel of the TV's user interface (UI) for ease of use and understanding.